skill-stocktake

To audit Claude skills for quality assurance, run the /skill-stocktake command to evaluate commands against a checklist using AI holistic judgment. It reviews recently changed skills or performs a complete stocktake of all active skills.

A quick scan evaluates only recently changed skills for faster quality assurance, while a full skill stocktake provides a comprehensive review of all existing Claude commands to ensure total quality coverage.

Yes, you can target specific projects for skill evaluation. The audit process allows customizable targeting to restrict the quality checklist and AI holistic judgment to only evaluate skills within your designated project scope.

To run a comprehensive review of all Claude commands, execute the /skill-stocktake command with the full option. This initiates a complete stocktake that audits every skill against a strict quality checklist using AI holistic judgment.

Yes, you need scripts to execute the Claude skill audit and cache results. The skill stocktake requires these execution scripts to process quality assurance checks and manage the evaluation output effectively.

A limitation of using AI holistic judgment for skill management is that it requires execution scripts to run and cache results. The audit relies on this structured checklist approach rather than manual code review to evaluate quality.